2
$\begingroup$

Let $f\in C_1\left(\mathbb{R}^n,\mathbb{R}\right)$, is it true that if $[a,b]$ is a closed interval and $\left(x_2,...,x_n\right)$ is fix, then for all $\varepsilon>0$ there exists $\delta$ such that: $$ \left|\frac{f\left(x_1+\delta,x_2,...,x_n\right)-f\left(x_1,x_2,...,x_n\right)}{\delta}-\frac{\partial}{\partial x_1}f\left(x_1,x_2,...,x_n\right)\right|<\varepsilon $$ for all $x_1\in[a,b]$. If yes, how to prove it? If no, are there similar but weaker statements which are true? Or are there even stronger ones?

In a way this expresses uniform convergence of the partial derivatives. But I wasn't able to remove the dependency on $x_1$ which is exactly the purpose of this assertion.

$\endgroup$
0
2
$\begingroup$

This result is true. Let $g:\mathbb{R}\rightarrow\mathbb{R}$ be defined by $g(x)=\frac{\partial}{\partial x_1}f(x_1,x_2,...,x_n)$. $g$ is continuous and hence $g$ restricted to the compact set $[a,b+1]$ is uniformly continuous. Therefore, given $\epsilon >0$, $\exists 1>\delta >0$ such that $\forall y,z\in[a,b+1]$, $|y-z|<\delta \implies |g(y)-g(z)|< \epsilon$. Now, $\forall x_1\in[a,b]$, by the legrange mean value theorem, we have $$\left|\frac{f\left(x_1+\delta,x_2,...,x_n\right)-f\left(x_1,x_2,...,x_n\right)}{\delta}-\frac{\partial}{\partial x_1}f\left(x_1,x_2,...,x_n\right)\right|$$ $$=\left|\frac{\partial}{\partial x_1}f\left(x_1+\gamma,x_2,...,x_n\right)-\frac{\partial}{\partial x_1}f(x_1,x_2,...,x_n)\right| (where\gamma\in(0,\delta))$$ $$=|g(x_1+\gamma)-g(x_1)|$$ $$<\epsilon$$

$\endgroup$
1
$\begingroup$

It can be proofed in a closed interval though this problem doesn't have to be a partial derivative because all other variables are fixed.

problem: let $f\in C_1\left(\mathbb{R},\mathbb{R}\right)$ and $[a,b]$ is a closed interval. For each $\epsilon \gt 0$, there exists a $\delta$ s.t if $|x-x_1| \lt \delta$, $|\frac {f(x)-f(x_1)}{x-x_1}-\frac{df(x_1)}{dx}| \lt \epsilon$ for all $x_1 \in [a,b]$

We understand that, by definition, there exists such a $\delta$ for each $x_1$ or $\delta(x)$ is defined in closed interval $[a,b]$. Our question is not changed to if there exists a $\delta$ s.t. $\delta(x) \ge \delta$ in interval $[a,b]$ This is true because the function is a $C_1$ function.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.